Trauma exposure significantly impacts mental health, with a substantial portion of individuals developing psychopathologies such as PTSD, PD, GAD, and MDD. These disorders share overlapping symptoms and biological underpinnings, highlighting the need to understand shared neural mechanisms beyond traditional diagnostic categories. Prior research has largely focused on PTSD, neglecting the neural signatures of trauma exposure in healthy individuals and those with various psychopathologies. This study aimed to address this gap by investigating the behavioral and neural markers of trauma and resilience using fMRI. Excessive threat generalization, where fear is inappropriately transferred from a dangerous stimulus to similar but safe stimuli, has been proposed as an endophenotype for several psychopathologies. Brain areas implicated in threat generalization and discrimination include the hippocampus, vPFC, insula, dmPFC, ACC, and thalamus, and associated networks like DMN, SN, and ECN. Previous studies have shown stronger generalization in psychiatric patients, but the neural markers in trauma-exposed individuals with or without psychopathology, and the temporal dynamics of threat discrimination learning, remain understudied. This study used a two-stage generalization fMRI task to examine behavioral risk ratings and network connectivity (DMN, SN, ECN) in trauma-exposed individuals with and without psychopathology, compared to healthy controls. The hypotheses were that trauma-exposed individuals with psychopathology would exhibit elevated generalization and limited changes in SN and ECN connectivity, compared to resilient trauma-exposed individuals and healthy controls.

The literature review extensively cites previous research on threat generalization as a potential endophenotype in several anxiety disorders, including PTSD, PD, and GAD. Studies have identified brain regions and networks involved in threat generalization and discrimination, such as the hippocampus, ventral prefrontal cortex (vPFC), insula, dorsomedial prefrontal cortex (dmPFC), anterior cingulate cortex (ACC), thalamus, and the default mode network (DMN), salience network (SN), and executive control network (ECN). However, a significant gap highlighted is the lack of research examining these neural markers in trauma-exposed individuals both with and without significant psychopathology, and a lack of research into the temporal aspects of threat discrimination learning, a crucial aspect of understanding how psychopathology develops and is maintained.

The study comprised 114 participants (62 trauma-exposed, 26 healthy controls). Trauma-exposed participants were further categorized into those with psychopathology (TEPG, n=31) and those without (TEHC, n=31). A generalization/discrimination fMRI task was used, consisting of three phases: pre-acquisition, acquisition, and generalization (with early and late stages). Participants rated their perceived shock risk (0-2) after each trial. fMRI data was acquired using two 3T GE scanners. Preprocessing included realignment, slice-time correction, outlier detection (ART), CompCor, spatial normalization, reslicing, and smoothing. Group ICA (GIFT toolbox) was used to identify intrinsic connectivity networks (ICNs), including SN, LECN, RECN, a-DMN, and p-DMN. Statistical analysis involved paired t-tests for comparing CS+ vs. CS- ratings, repeated measures ANOVAs for comparing behavioral and neural changes across generalization stages, and one-way ANOVAs for comparing LDS changes across groups. Linear deviation scores (LDS) measured the steepness of generalization gradients, with higher values indicating stronger generalization. Covariates included scanner type, age, and sex.

Comparing trauma-exposed (TE) and healthy controls (HC), the study found that TE participants demonstrated lower activity reduction in the SN and RECN across the two generalization stages, and significantly worse discrimination learning in the SN as measured by LDS. When comparing TEHC, TEPG, and HC, TEPG displayed consistently higher risk ratings across both generalization stages compared to TEHC and HC, indicating poorer discrimination. TEHC, unlike TEPG and HC, maintained higher SN and RECN activity over the two stages, suggesting a resilience mechanism where sustained SN activity may work in conjunction with RECN activity to improve discrimination. In HC, a higher reduction in SN activity and better discrimination learning (LDS) was observed over time compared to TE. The findings demonstrate that trauma exposure maintains high SN and RECN activity over time, hindering the ability to use the SN effectively for stimulus discrimination. Resilient trauma-exposed individuals showed a compensatory mechanism, higher RECN activity associated with discrimination, which was not observed in trauma-exposed individuals with psychopathology.

The findings support the hypothesis that trauma exposure leads to distinct behavioral and neural markers. The reduced activity reduction in the SN and RECN across the two stages and poorer discrimination learning in SN suggests a trauma exposure phenotype impacting threat processing and discrimination. The maintained high SN activity coupled with enhanced RECN activity involved in discrimination in resilient individuals suggests a resilience signature characterized by sustained attention and adaptive cognitive control. The findings extend previous research by demonstrating that trauma exposure itself does not invariably impair discrimination, but rather this deficit is linked to the development of psychopathology. The results emphasize the importance of targeting specific neural dysfunctions in trauma-exposed individuals to improve treatment efficacy, with potential interventions focusing on enhancing the efficiency of the SN and RECN in threat discrimination and response regulation.

This study provides evidence for both trauma exposure and resilience signatures in neural activity and connectivity following trauma. The findings highlight the SN and RECN as key circuits involved in threat generalization and discrimination. The ability of resilient individuals to utilize higher RECN activity for discrimination might represent a compensatory mechanism. Future research should explore longitudinal studies to determine the causal relationship between these neural signatures and the development of psychopathology. Interventions targeting the SN and RECN, such as tDCS, hold promise for improving treatment outcomes in trauma-exposed individuals.

The study has several limitations. The lack of limbic area differences (e.g., hippocampus) may be due to the task design not involving context learning. The use of two scanners, although addressed through data harmonization, may introduce noise. The exclusion of individuals with certain diagnoses may limit generalizability. The cross-sectional nature of the study prevents inferences about causality, requiring future longitudinal studies to clarify the temporal relationship between neural signatures and psychopathology development.